My battle with LED attenuation

[Paul Barker]

sub sectioned this out in case God forbid anyone is not interested in my PP amp thread.

This morning I have been annoyed by the channel imbalance I didn't worry about in the excitement of yesterday.

So until I can afford to order a pile of NSL-32SR2S's then make up a rig to match them on, I have decided to add the (1k to 5k) pots suggested in the Lightspeed diagram below. All these do is vary the voltage so they are not in the signal path. Undoubtedly then I shall be able to recover balance, though it will quite likely need a tweek at different voltage levels. For me I think this would be acceptable.



Adendum, it appears that the optocouplers I ordered from Farnel were not the S (matched) postcript but they are . (dot) postscript.

RS do the S postricpt but you pay an additional £10 per order, and you probably still need to buy enough to match them.

I have been researching dual output optocouplers in anticipation dual output would be matched. But available from UK soil they are thin on the ground and costly.

So far another pot to twidle looks likely outcome. that or order a ton of cheap ones and match up. but will they drift over time anyway? A search reveals people keep sending their OEM products back when mismatched, they are sent out again matched and later yep, back they go again.

New learning curve!

Of course, it is obvious, how come nobody else posted this? Because they prefer to sell you a matching service?



Centre the pot for your preferred listening position and preferred listening level. Maybe on occasional alternative listening levels, or if you have one of those rogue amps with channel imbalance you can't be R'sd to fix, get up and give the balance control a tweek!

the wander of it is, we bring that that well useful thing of history the balance control which we formerly through away to simplify our audiofool signal path. well here it is back, you at last have control again, but not in the audiofoolery coveted signal path.

elegant or what?

If you are a total nut job like many of the audiofoolishnuts on the web you could build your own log stepped attenuator to replace the log pot and find the exact value of series resistance for the one channel <and that may switch channels at different levels> which requires it for balance at every single attenuation level.

Think that is mad? Try paying for matched LED's instead? Who is more mad? My way cheaper.

[Paul Barker]

goes without saying there is a input to ground complement, but balance should need only adjustment on the series element, as that would equally balance any variance in shunt resistance to the point where it matters. If you think it will affect your aural pleasure please be my guest and over complicate it.

[JamesD]

Neat Stuff!!! And your balance control is great, just what I need as my hearing is fading in my right ear at a different rate than my left ear

Many thanks for finding and posting this here.

J

[Paul Barker]

OK I stuck in a 4k7 trimmer wiper to output from pot each leg to each series LED. Yes I have restored balance, but it has taken the entire 4k7 over on one channel. Simples, a 4k7 resistor is put in one side, and a pot is connected again as shown.

in other words gross adjustment by fixed resistor, balance by pot.

[Nick]

Or just use two volume controls

I built one of these some time ago. Matching the LDR's is a major pain and a hiding to nothing, as the response curve varies wildly between samples. The best you can hope for (as Paul found) is match at a single point. but as there are two in each chan and you dont know what point they need matching at...

The only way I could see of getting close matching is by using a curve tracer.

There is loads of talk on the web about the need for matching. I did toy with the idea of wrapping they with a set of relays, an A to D converter, and four D to A converters and micro to drive them so the characteristic could be learnt after building and the controller driving them to match. The solution started to get worst than the problem.

[IslandPink]

So these are light-dependant resistors, but the light is enclosed in the package - and provided by an LED whose output you set by altering the current through it ? Not seen this before, just trying to get my head around the concept .

[andrew Ivimey]

I remember when these came out. Back in my day sonny we used to scratch the paint off an 0C71 and use a light bulb!

I'm interested in this way of controlling the volume.

[Andrew]

. I did toy with the idea of wrapping they with a set of relays, an A to D converter, and four D to A converters and micro to drive them so the characteristic could be learnt after building and the controller driving them to match. The solution started to get worst than the problem.

Kinda like what modems used to do with active eq, so that they could deal with different line characteristics. It would be a great solution to the problem, but are these things so much better than the alteratives that its worth doing?

Andrew

[Nick]

but are these things so much better than the alteratives that its worth doing?
Andrew

The one I built sounded ok, but not any better. A stepped attenuator sounded better IMHO. The other problem is the LDR resistive element is not entirely linear. But I never measured them to see if that was significant. It would be interesting to measure Pauls at Owston.

[Paul Barker]

Mine does sound better, very clear, makes the panasonic pot sound smeared with film.

I have even begun considering using LDR's as resistors, but it would entail a host of power supplies all over the place!

as Nick says a tweek is needed at each volume level, but it isn't grossly out once the primary imbalance is adjusted. In my case I have to get up to adjust volume anyway. Regarding the dual channel pots I have never got on with those. I prefer a balance pot, and only decided against it to keep the signal path cleaner. I think that is all dual pots came out for.

Mark the perceived benefit is no surface tension for the signal to cross as there is across the wiper. Whereas when you adjust a pot to control the LED you are doing it at higher current than the audio signal so it can traverse any resistances unaffected, and in any case it isn't in the signal path.

witth a stepped attenuator you have resistor quality to consider and also this junction switch to get the miniscule signal current across. My stepped attenuator was very nasty, took me no time to bin it, less time that the Sowter transformer got.

[Mike H]

Re the LDR's, lord I wouldn't, for the reasons you've already listed. I did once work out a theoretical way of using a control Voltage with op-amp feedback as a 'servo control' to keep them in balance, but needs extra LDR's for the op-amp gain feedback control, and which again must be well matched to their companion signal LDR's anyway. The idea was that an op-amp, given some sort of Volume level control input, produces the equivalent resistance of the LDR by looking at the resistance of the control LDR, and altering the current to the LED's or whatever as is required. Needs 2 matched LDR's. I was thinking in terms of a bulb in the centre of a tube with a LDR inserted into each end. Seriously. One LDR does the audio signal level, the other is the op-amp level control reference. The idea then started getting very compicated if you then want TWO signal LDR's, the one in series with the path of the signal and the other in parallel to it as the attenuator part.

[Paul Barker]

Well in the event the balance which I adjusted this AM at lowish level has held throughout all my listening range.

[Nick]

Re the LDR's, lord I wouldn't, for the reasons you've already listed. I did once work out a theoretical way of using a control Voltage with op-amp feedback as a 'servo control' to keep them in balance, but needs extra LDR's for the op-amp gain feedback control, and which again must be well matched to their companion signal LDR's anyway. The idea was that an op-amp, given some sort of Volume level control input, produces the equivalent resistance of the LDR by looking at the resistance of the control LDR, and altering the current to the LED's or whatever as is required. Needs 2 matched LDR's. I was thinking in terms of a bulb in the centre of a tube with a LDR inserted into each end. Seriously. One LDR does the audio signal level, the other is the op-amp level control reference. The idea then started getting very compicated if you then want TWO signal LDR's, the one in series with the path of the signal and the other in parallel to it as the attenuator part.

Yep, but sadly, it doesn't solve anything. For it two work, you need two perfectly matched LDR's if you have two perfectly matched LDR's then you just use them, you dont need the additional complexity. But the problem is you can't get two perfectly matched LDR's.

One solution is to add a HF signal to the input, and filter that at the output, but monitor the level of that HF at the output and use that to close the loop with a servo and balance that way. Again, seems to me a solution thats worst that the problem.

[JamesD]

My hearing has become asymmetrical and the asymmetry changes with frequency and level... as a consequence I find the balance is different for each piece of music I listen too... so a good balance pot or separate volume controls on each channel works very well for me...

The nice thing is that when the magic happens things like balance settings are irrelevant!

J

[Mike H]

Paul, yes exactly